Multiple Discourse Analysis

8/9/2019 Multiple Discourse Analysis

1/38

A B S T R A C T This article explores the contributions that ve differentapproaches to discourse analysis can make to interpreting and understandingthe same piece of data. Conversation analysis, interactional sociolinguistics,politeness theory, critical discourse analysis, and discursive psychology are theapproaches chosen for comparison. The data is a nine-minute audio recordingof a spontaneous workplace interaction. The analyses are compared, and thetheoretical and methodological implications of the different approaches arediscussed.

K EY W OR DS: conversation analysis, critical discourse analysis, discourse analysis,discursive psychology, interactional sociolinguistics, politeness theory, pragmatics,workplace interaction

Any newcomer to the study of conversation or language in use will be bewil-dered by the array of analytic approaches that exists. Even more seasonedresearchers might be challenged to provide comprehensive descriptions of therange of discourse analytic approaches available in disciplines across thehumanities and social sciences. These include pragmatics, speech act theory,variation analysis, communication accommodation theory, systemic-functionallinguistics, semiotics, proxemics, and various types of rhetorical, stylistic, seman-tic and narrative analysis. A recent interdisciplinary textbook (Titscher et al.,2000), for example, surveys 12 different approaches to discourse analysis, andeven then three of the ve approaches adopted in this article are not included.These ve approaches to the analysis of spoken interaction will be well known toreaders of journals such as this one, but we make no claims for comprehensive-ness here. Rather, our aim in this article is to explore the different facets of oneparticular spoken interaction by providing a detailed discourse analysis of its fea-tures from ve different analytical perspectives.

A RT I C L E 351

Multiple discourse analyses of aworkplace interaction 1

M A R I A S T U B B E , C H R I S L A N E, J O H I L D E R,E L A I N E V I N E , B E R N A D E T T E V I N E , M E R E D I T H M A R R A, J A N E T H O L M E S , A N D A N N W E AT H E R A L L 2V I CT O RI A U N IV E RS I TY O F W E LL I NG TO N

Discourse StudiesCopyright 2003

SAGE Publications.(London, Thousand Oaks,

CA and New Delhi)www.sagepublications.com

Vol 5(3): 351388.[1461-4456

(200308) 5:3;351388; 034260]
http://www.sagepblications.com/http://www.sagepblications.com/


2/38

Most textbooks (including Schiffrin, 1994; Paltridge, 2000; Titscher et al.,2000) tend to focus on the contrasting theoretical and conceptual features of dif-ferent approaches, emphasizing their very different objectives, and the conse-quences for what their practitioners regard as relevant in their analyses. Whilesome textbooks do exemplify how particular approaches work in practice in rela-tion to particular textual materials, we are not aware of any which attempt tosystematically apply a series of different approaches to the same naturally occur-ring spoken material. At the time of writing (1999), the members of ourDiscourse Analysis Group shared interests in sociolinguistic, social psychologicaland critical perspectives on language in use. Approaches in the forefront forgroup members in their own research were one or more of conversation analysis(CA), interactional sociolinguistics (IS), pragmatics/politeness theory, critical dis-

course analysis (CDA) and discursive psychology. Most group members describedtaking an eclectic approach to their analyses, applying elements of one or moremodels as relevant to their research objectives.

In this article, we identify the core elements of each of these ve approaches;we highlight the key insights into the data provided by each analysis; and weidentify the respects in which they overlap or offer distinctive perspectives on thedata. The analyses are presented in an order approximately reecting their posi-tion along a continuum from micro-analysis to macro-analysis of the sampleinteraction. It should be noted that the individuals applying each sample analysisare not necessarily experts in that approach, nor do they necessarily subscribefully or solely to it themselves. The analyses are exemplary rather than denitive.We hope, however, that they will provide a useful complement to the more theo-

retical surveys of methods of discourse analysis provided by textbooks such asTitscher et al. (2000).

The dataThe interaction analysed below comes from tape-recorded data collected as partof the Language in the Workplace project at Victoria University. 3 It is a nine-minute excerpt relating to a single issue, taken from the beginning of a longermeeting between two people in a New Zealand workplace. Further contextualand ethnographic information will be provided as it becomes relevant to eachanalysis. A full transcript of the interaction appears in Appendix 2. 4

Our starting point was an existing fairly broad orthographic transcription of

the extract based on the conventions used in the Wellington Corpus of SpokenNew Zealand English (see Appendix 1), with utterance numbers added for ease of reference. This provided the group with a convenient baseline version of thetranscript to use in conjunction with the audio recording when analysing anddiscussing the excerpt, and also one which was reasonably neutral in that it wasnot linked to any particular analytic approach. This is the version reproducedin this article. However, this solution to the question of which system(s) of transcription the group should use immediately raised a number of interesting

352 Discourse Studies 5(3)


3/38

theoretical and methodological issues. Any transcription can of course only everbe an approximate and partial rendition of the recording on which it is based,and different approaches to discourse analysis have developed their own tran-scription systems precisely because they wish to address specic research ques-tions which require varying degrees and types of detail in the way the data isrepresented on paper.

For example, some approaches, such as CA and IS, involve a micro-analyticinvestigation of interaction sequences, which is typically reected in a more ne-grained and detailed transcription system than is found in other forms of dis-course analysis. These transcripts typically provide a precise marking of prosodicfeatures such as pause length, changes in pitch contour and a range of vocaliza-tions, as well as an accurate rendition of aspects of the turntaking system such

as overlapping speech, feedback and latching. While such a transcription is com-plex and time-consuming to prepare, because the analysis is similarly intensive, itis usually restricted to relatively short sequences of interaction. Conversely, otherapproaches such as CDA and discursive psychology generally require the analystto range over far more extensive samples of text to identify macro-discoursal pat-terns. Although this does not exclude some detailed analysis of short excerpts, onthe whole, these approaches require a far less detailed transcription of the data.

Exploring these differences highlighted for us the extent to which no tran-scription can ever be a neutral or complete rendition of a spoken text. Theprocess of transcription is inevitably selective, and therefore involves a certainamount of interpretation and analysis. Moreover, the aspects of a piece of datawhich are represented in a given transcript, and even the way it is set out on the

page (cf. Edelsky, 1981), will affect what we notice about it, how we interpret it,and what aspects we are most readily able to take account of in our analysis.Although an in-depth discussion of the complex relationship between a particu-lar transcription and the spoken data on which it is based is beyond the scope of the present article, the group did take explicit account of this issue. In preparingthe CA and interactional sociolinguistics analyses, for example, relevant parts of the extract were re-transcribed using a ner-grained CA-style transcription, andeach analysis made use of the actual recording, not a transcript alone. However,for ease of reference and consistency, we have opted to reproduce the extract andany examples cited throughout this article in a generic format.

1. Conversation analysis (Chris Lane and Jo Hilder)

Conversation analysis (CA) is a sociological approach to the analysis of interac-tion. Harvey Sacks, the founder of CA (Sacks, 1992), saw it as a basis for anobservational science of society. Useful introductions to CA are provided byLevinson (1983), Heritage (1984), Psathas (1995), Pomerantz and Fehr (1997),Hutchby and Wooftt (1998), and ten Have (1999). CA focuses on analysis of recordings of spontaneous spoken interaction (formal as well as informal, insti-tutional as well as private/personal).

Stubbe et al.: Multiple discourse analyses of a workplace interaction 353


4/38

CA is based fundamentally on a model of communication as joint activity(Sacks, 1984), like dancing or joint musical performance. It rejects the typicallinguistic model of communication as sending and receiving messages. In otherwords, it says that dialogue is not a succession of monologues. So, CA is con-cerned with how the jointly organized activity of talk-in-interaction is carriedout, and how participants produce joint achievements such as conversationalclosings, storytelling, disputes, medical diagnosis, the mutually dependent rolesof interviewer and interviewee, and so on.

Sequences are an important focus of an analysis and each utterance (or ges-ture) is understood as a step (action) in a joint activity. Thus one of the mainfocuses of CA is on how interaction unfolds across sequences of actions by differ-ent participants. The signicance of an utterance or gesture is highly dependent

on its position in a sequence, as well as being jointly negotiated, and this is onereason for conversation analysts reluctance to aggregate instances of utterancetypes for quantitative analysis (Schegloff, 1993; Wieder, 1993). CA thus differsfrom approaches which (typically) analyse one utterance at a time, such asspeech act theory and politeness theory.

The unfolding of interaction depends on the interpretation of a currentspeakers utterance by the next or a subsequent speaker, and to show that theyare engaged in a joint activity, they need to display that interpretation in someway. Even if the next speakers interpretation is wrong from the originalspeakers point of view, it is open to the original speaker to offer a correction. Ingeneral, any utterance can be interpreted in numerous ways by analysts; for CAit is important to nd evidence in the interaction of which of these possible inter-

pretations have been taken by the participants. The following is an example fromthe extract of how the unfolding of action depends on the participants displayedorientation to the interaction:

absolutely nothing + sinister or any other agenda other than thatno im not looking for that

At line #62 there is a spontaneous denial of bias by Tom, which displays hisunderstanding that Claires complaint may be one of discrimination on somegrounds; Claire appears to deny this in #64.

Participants actions in interaction are (in general) done as locally occa-sioned, for example, as responsive to a prior utterance, or as relevant to a currentnon-linguistic activity, so that they can be seen to t into a current sequence of

actions (again a problem for quantifying across contexts). Spontaneous interac-tion thus has an improvised character.Functional categories of utterances are not based on analysts attempting to

read speakers intentions, but rather on their responsiveness to earlier actionsand on the actual or potential following actions: in speech act theory terms, CA ismore concerned with perlocutions than with illocutions. Possible functions of utterances are not specied in advance (as in speech act theory); rather, newfunctional categories are found as different aspects of interaction are studied.



5/38

The notion of normative rules is another feature of this analytic approach. InCA, rules (e.g. for turn-taking, for how sequences can unfold) are not invariantdescriptive rules in the linguistic sense, or statistical generalizations, but rathernormative and interpretative they provide a reference point for participants totreat actions as unremarkable or deviant; participants justify actions as followingshared rules or as accountably violating such rules, complain about others vio-lations, apologize for their own violations, etc. Such references to rules are notnecessarily consistent across different individuals or across different interactionsbecause they are locally occasioned. These rules generally operate below the levelof consciousness or awareness, as we see in this example:

can i just grab th- just grab that phonesorry about that

thats okay:On the basis of their wordings, these look like a request, apology and an accept-ance, and the sequence is indicative that Toms answering the phone violatessome rule. On the basis of this one example we can hypothesize a rule that aninteraction once started should be continued (through to a negotiated close)rather than abruptly stopped to start another interaction: whether this is a gen-eral rule could be tested not experimentally but by examining other recordeddata for evidence of such a rule in operation.

CA also takes an approach to understanding context that differentiates itfrom many other approaches. Context is not seen as given prior to interaction.Social and contextual factors such as participants identities are not analysed asindependently speciable causes of behaviour, but rather as resources thatcan be invoked as relevant in a normative/interpretative way, or in fact contex-tual factors can be constituted by the interaction itself, as the following exampleillustrates.

yeah um yeah i want to talk to you about um oh its a personal issue um + well i-the decision to make um jared acting manager while joseph is away

In #1, Claires formulation its a personal issue paradoxically indicates that thisis an institutional interaction of some kind because it implicates that it is rele-vant to make a distinction between personal and non-personal issues.

well ive been overlooked quite a few times but i wanted to nd out specicallyhow what i could do to help myself be considered next time

In #5, Claires apparent request for advice implicates an advising or mentoringrole for Tom and an advisee role for Claire. In #81, because hes your immediatecontrolling ofcer is an instance of invoking a particular contextual factor,namely the workplace hierarchy, in order to accomplish a particular situatedaction, i.e. redirecting Claires request for advice to Joseph.

Thus, there is a strong CA position on context the interaction is the context.Analysts do not need to look outside the interaction unless some external factoris invoked in the interaction. However, there can be a weaker and more exible



6/38

position. Participants need to display to each other their (degree of) mutualunderstanding, and what context is relevant for them; but what is enough forparticipants may not be enough for analysts, and we may need additional help tointerpret what the participants are saying/doing.

Aside from context, CA also has quite a unique position on model-building.CA is very circumspect about premature theorizing and formalisation, and hasconcentrated on rather microscopic empirical studies of specic aspects of inter-action. CA has accumulated a large number of detailed studies of interactionalactivities, sequences, and the uses of particular devices such as code-switching,particles like OK and oh, and various intonation contours (Couper-Kuhlen andSelting, 1996), mainly in English, but increasingly now in other languages. Theaccumulated set of studies is beginning to form a fairly integrated model of how

talk-in-interaction is conducted as a joint activity.

I L L U S TR ATI V E A N A LY S I S (1 ) O F I N TE R AC T IO NMany CA analyses are possible of this material, depending on what phenomenayou pay attention to, just as there would be many possible phonological analyses.So we can only explore some of what CA has to offer. We concentrate here onClaires presentation of the topic/problem and Toms response to that:

yeah um yeah i want to talk to you about um oh its a personal issue um + well i-the decision to make um jared acting manager while joseph is away

We do not know what comes before this but the utterance itself has the characterof a topic initial (i want to talk to you about) which indicates that this will be an

extended turn, together with a formulation of the topic (its a personal issue)and an initial characterization of the topic (the decision . . . away).The continuers mm and mm by Tom are evidence

(though not denitive) that Tom hears Claires presentation of the topic asincomplete, and is returning the oor to her to complete it. At #3 he could havetreated #1 as a complaint or some such and responded to it in those terms, butthe fact that he does not allows #1 to stand as a preface to Claires presentation of her concerns:

and i wanted to get somewell ive been overlooked quite a few times but i wanted to nd out specically

how what i could do to help myself be considered next time

These put forward issues or concerns which can be heard (in a taken-for-granted way) as related to the topic initiated in #1, on the basis that Claire is fol-lowing the normative rules of turn-taking and topic initiation. Utterance #5looks initially, from an analysts point of view, like a generalized complaint (fromwhich a specic complaint about the decision referred to in #1 could be inferred)and a request for advice or guidance. But it remains to be seen if these analystinterpretations are actually oriented to by the participants themselves.



7/38

(well) i just want to talk to you about it and and i suppose [swallows] [tut]i just want to get some ideas on what i could do to actually be consideredfavourably next time

This restates the topic Claire began in #1, #4 and #5, partly by using topic initialtalk (I just want to talk to you about . . .), partly by paraphrasing part of #5,namely the part that looked like a request for advice/guidance, partly by thepronominal reference it. With these devices Claire in effect returns the interac-tion to the point reached at the end of #5.

yeah i dont think its a its a question of er favourabilityi mean it was a question more of practicalities more than anything elseum i was urgent need of someone to ll in and jared had done that in the past

already

so i mean there would be very little chance of me crossing paths with the p mthe policy manager you know having gone for someone that hes [voc] for hisreasons um hes er had um sitting in a position

plus the fact that i suppose it had a little bit also to do that with the fact that era number of very current issues that i had been involved with jared on er like[name of organization] + and like the [topic] er issue

um i probably had more immediate contact with him you know(and further utterances through to #43)

This is Toms opportunity to display his interpretation of Claires contributions in#1, #4, #5 and #16. Even though she has nished with something that lookslike a request for advice, and there is a preference for responding (at least ini-tially) to the most recent contribution, Toms turn here displays an interpretation

of her contributions as a complaint about the decision to appoint Jared, andappears to be a justication of that decision and hence a rejection of the com-plaint. It appears that #17 is an immediate, tied response to #16, through thelexical relation of favourably and favourability. This device, which appearsto exploit the ambiguity of these words, provides an improvised bridge to Toms justication. It resembles techniques used by interviewees in broadcast inter-views to reformulate questions before providing a response to the (reformulated)question.

Both Claires and Toms contributions look like what in CA are called dispre-ferred actions, because of characteristics like hesitation, delay, qualifying (hedg-ing) and indirection. In CA these are features which mark actions or sequencesas problematic in respect of interactional norms: it could be the specic

complaint-justication sequence which is dispreferred here, or the more generalproblem of dealing with a sensitive topic.Toms justication extends from #17 through to #43. Through this passage,

Claire provides verbal continuers and response tokens: yeah, right, mm, oh,okay. On reading the transcript these responses look like acceptances of Toms justication. But on listening to the tape these tokens are noticeably low pitchedand in most cases drawled with a level tone and at a low volume. This prosodycould be described as sounding neutral, unenthusiastic or resigned, but



8/38

whatever the description, it strongly marks these responses as not being positiveacceptances. This withholding of acceptance may account for the way Tom keepsextending his justication and in fact repeating components of his justication inparaphrased form. These response tokens can be contrasted with those at the endof the excerpt, which sound distinctly positive and enthusiastic.

In summary, CA provides a method for analysing in detail the way partici-pants jointly construct the interaction and at the same time constitute thecontext, including participants identities. With interactional sociolinguisticsit shares an interest in the process of contextualization, and with discursivepsychology an interest in accounting and justifying practices.

2. Interactional sociolinguistics (Elaine Vine and Maria Stubbe)Interactional sociolinguistics (IS) has its roots in the ethnography of communi-cation, and analysts using this approach typically focus on linguistic and cul-tural diversity in communication, and how this impacts on the relationshipsbetween different groups in society. IS also represents an approach to discourseanalysis which attempts to bridge the gap between top-down theoreticalapproaches which privilege macro-societal conditions in accounting for com-municative practices, and those, such as CA which provide a bottom-up socialconstructivist account (Gumperz, 1999: 4534). IS draws heavily on CA tech-niques in its microanalytic approach to interactions, but unlike CA, an IS analy-sis explicitly recognizes the wider sociocultural context impacting oninteractions. John Gumperz, generally regarded as the founder of interactional

sociolinguistics, characterizes the approach as building from a conversationanalysis approach in the following way:

we must turn to a speaker-oriented perspective and ask what it is speakers and listen-ers must know or do in order to be able to take part in a conversation or to create andsustain conversational involvement. By formulating the basic issues in this way, thefocus shifts from the analysis of conversational forms or sequential patterns as suchto the necessarily goal-oriented interpretive processes that underlie their production.(Gumperz, 1992: 306)

Interactional sociolinguistics taps into those goal-oriented interpretiveprocesses through what Gumperz calls contextualization cues. These are con-stellations of surface features of message form . . . by which speakers signal andlisteners interpret what the activity is, how semantic content is to be understood

and how each sentence relates to what precedes or follows (Gumperz, 1982:131). Contextualization cues relate to contextual presuppositions (tacit aware-ness of meaningfulness) which in turn allow participants to make situated infer-ences about the most likely interpretation of an utterance. Speakers can makechoices between features at any of a number of levels including: (i) code, dialector style; (ii) prosodic features; (iii) lexical and syntactic options, formulaic expres-sions; and (iv) conversational openings, closings and sequencing strategies(Gumperz, 1982: 131). Non-verbal behaviours also provide very important



9/38

contextualization cues (e.g. Schiffrin, 1996), although obviously these are notavailable for analysis in the case of audiotaped data.Roberts et al. (1992) suggest that IS provides a useful tool for the critical

analysis of discourse in certain typical strategic research sites. These are situa-tions of public negotiation such as interviews, meetings, and encounters atwork, which are characterized by status and power differentials between the par-ticipants. There is often tension between different (and sometimes conicting)goals, which is played out through the discourse processes we can observe.Information about the sociocultural context can throw further light on why cer-tain linguistic features are chosen and how these are interpreted by participantsin such contexts.

Based on the background information we have available about the sample

interaction, this is an excellent example of a strategic research site for an ISanalysis. The discussion takes place between a senior public service manager,Tom, and an analyst, Claire, who is two ranks below him in the organizationalhierarchy. From the ethnographic eldwork that was done at the time of the datacollection, we know that Claire is annoyed that she was overlooked for the sharedacting manager position she believes she was promised by her own manager, andthat she and some of her female colleagues interpret this as another example of gender discrimination within the organization. We also know that she hasexpressed her intention to raise the issue with Tom, because regardless of whether or not he was ultimately responsible for changing the earlier decision,he has the power to inuence what will happen in future.

I L L U S TR ATI V E A N A LY S I S (2 ) O F I N TE R AC T IO NWe only have space here to focus on selected aspects of the sample interaction,but this should be sufcient to provide an indication of what an IS analysis has tooffer. We begin with an analysis of the contextualization cues provided by Clairein her opening utterances and how these are responded to by Tom.

In #1, Claires rst yeah has a falling tone and sounds like a response to aprevious turn. The second yeah has a rising tone and sounds like the beginningof something new. The phrase which follows, . . . i want to talk to you about umoh its a personal issue , has high pitch and there is a very high rise on personalissue and a noticeable stress on personal, all of which contributes to making thespeaker sound nervous and/or tense. There is also a hesitation before um ohwhich the original transcriber has not judged to be a pause, but which con-

tributes to the sense that the speaker sounds nervous. The speaker then contin-ues with um and a pause, and a false start well i - before stating the issue thedecision to make um jared acting manager while joseph is away. The hesitationsand false start contribute again to an impression that the speaker is nervous.

We can infer that this speakers nervousness and/or tension may arise fromseveral aspects of the sociocultural context. First, the speaker is lower in theworkplace hierarchy than the person she is bringing her issue to, and by doing soshe is, at least potentially, being critical of her superiors judgement. Second, by



10/38


11/38

it. We have already noted that his choice to follow up the complaint purpose mayhave led to hedging, but he now has two more reasons to hedge Claire is sound-ing unconvinced, and Tom has grabbed the oor through his use of latching. Wedo not have space to analyse Toms use of hedging in detail, but we would like tonote one aspect of it. He begins #21 with another speaker-oriented pragmatichedge i mean, then uses an addressee-oriented hedge you know here andagain in #23. The addressee-oriented hedges imply a shared perspective and,indeed, they connect with backchannels from Claire which are not just a baremm: #24 right and #27 oh okay, which may give Tom encouragement thatthey are nally moving towards a shared perspective.

As we say above, in addition to prosodic and pragmatic choices, speakers alsomake lexicogrammatical choices which can function as contextualization cues.

We will demonstrate this in more detail by tracing how this type of cue is used byTom and Claire at different points in the interaction to negotiate where theresponsibility lies for the decision to make Jared acting manager.

When Claire rst introduces the issue she wishes to discuss in #1, she refersto the decision. She has chosen a nominalized form, rather than a verb formwhere she would have to specify a subject, e.g. you decided, and she has chosennot to specify the decider by using the denite article the decision, rather than apronominal form your decision. In this use of neutral agency, she is distancingherself from implied criticism of Tom. In #5, she goes further by implyingthrough her request for advice that she had done something wrong herself, andso could do better. Taken in isolation, Claires utterances clearly do not allow usto draw the inference that she is holding Tom in some way responsible for the

decision. However, this does become a reasonable inference once we take intoaccount the difference in status, our knowledge that Claire was angry about thesituation, and the evidence we have from later in the interaction (#1234)where she makes it clear that she had indeed been promised the position by Joseph. This interpretation is supported by the way in which Tom clearly orientsto the implied criticism of his actions in lines #1743, as already discussed above.

Tom appears to go along with distancing himself from the decision in #17and #18, also referring to the decision neutrally as it. Then in #19, he beginsusing the personal pronoun I, but with respect to his reasons (i was in urgentneed of someone), not the decision itself. It is not until #28 that Tom explicitlytakes ownership of it as my decision, and this is after he has given several rea-sons, and Claire has indicated through the more positive backchannels we noted

earlier that a shared perspective may be emerging.By #33, however, Tom is moving responsibility for the decision into Josephs

court: because Joseph did not communicate it clearly, Tom had to step into thebreach. In #40 Tom reiterates that it was his decision, but only acting on prece-dents in the absence of a clear communication from Joseph. In #59, Tom offersClaire some advice it was Josephs decision and she should address this as alonger-term issue with him and provides a number of reasons as to why hecannot intervene directly.



12/38

In #123 and #124 Claire reframes this once again by informing Tom that Joseph had in fact communicated to her that she and Jared would be sharing theposition. The inference is that either Tom did override Josephs decision, or hegenuinely did not know about it, and made a different decision himself for hisown reasons. This results in Tom accepting responsibility for the decision onceagain: #130 what I will do though in lieu of any decision coming to my atten-tion um make the decision. Claire is then in a position nally to seek an assur-ance in #133 that he will consider her next time and Tom nally gives Clairethe outcome she has been looking for:

so . . . I mean + next time it happens and if it does happen again then yeahsure no difculties.

all right then oh good

okay?okay thanksokay

To conclude, while this has been only a brief sample analysis of small parts of theinteraction, it does demonstrate how interactional sociolinguistics takes a micro-analytic approach within a wider sociocultural context. Unlike the CA analysisabove, an IS approach allows us to take explicit account of the unstated assump-tions and background knowledge the participants in an interaction bring to bearas part of the interpretive process. The analysis also assumes intentionality onthe part of both participants, and by tracing the way in which both speakers usevarious contextualization cues to frame their intended meanings, we have beenable to see how they move towards a shared interpretation by the time this part of their meeting comes to an end.

3. Politeness theory (Bernadette Vine and Maria Stubbe)

Politeness theory, as developed by Brown and Levinson (1978, 1987), has been amajor preoccupation of research in pragmatics for the past 25 years (Thomas,1995: 149). In Brown and Levinsons framework, linguistic politeness revolvesaround the concept of face (cf. Goffman, 1967). In particular, they identify aModel Person as someone who has two particular wants . . . the want to beunimpeded and the want to be approved of in certain respects (1987: 58).Brown and Levinson (1987) refer to these two wants as negative and positiveface respectively.

Certain kinds of speech acts, however, intrinsically threaten the face of thespeaker, hearer or both interactants (1987: 658), for example, requests, compli-ments and apologies. Brown and Levinson group these face-threatening actsaccording to whether they threaten the negative or positive face of the hearer orthe speaker. There is some overlap in their classication, as some face-threaten-ing acts (FTAs) threaten both positive and negative face and may threaten boththe speaker and the hearer.



13/38

A complaint or challenge is an FTA which threatens the positive face of thehearer by indicating that the speaker has a negative evaluation of some aspect of the hearers positive face. In making a complaint, the speaker indicates that theydo not like or want one or more of the hearers wants, acts, etc. In line #1 of theinteraction, for example, Claire indicates that she does not like Toms decision tomake Jared acting manager instead of Claire:


14/38

going off record. Her complaint is not directly stated; rather she tells Tomthat she wants to talk to him about a personal issue and asks for advice inrelation to her being overlooked for the acting manager position. In using anoff-record approach such as this, Brown and Levinson argue that the speakeris being as polite as they can be, apart from avoiding the FTA altogether. Theuse of this superstrategy affords the speaker the opportunity of evadingresponsibility for the FTA (by claiming if challenged that the interpretation of x as an FTA is wrong) and simultaneously allows S to avoid actually impos-ing the FTA x on H, since H himself [sic] must choose to interpret x as an FTArather than as some more trivial remark (1987: 73). We can see thishappening in Claires response to Toms summing up in line #64 of the inter-action, towards the end of his lengthy justication of the decision to appoint

Jared instead of Claire:cause Ive given you my reasons why I did it//abso\lutely nothing + sinister or any other agenda other than that/right\\ no Im not looking for that

5. Avoidance of FTA. The last superstrategy described by Brown and Levinson(1987) is not to do the FTA at all. This, of course, is the most polite super-strategy.

The second and third superstrategies earlier, use of positive and negative polite-ness, are realized in discourse by a variety of output strategies. These outputstrategies represent attempts by the speaker to address the face needs of the

hearer. Brown and Levinson claim that a persons face concerns can only be sus-tained by the actions of others (1987: 60), and that it is therefore in the interestof individuals to maintain each others face. For this reason, while they brieydiscuss and illustrate FTAs which damage the speakers own face needs (1987:678), Brown and Levinson do not include strategies which address these needsin their catalogue of negative and positive politeness output strategies. A numberof commentators have, with some justication, criticized this aspect of Brownand Levinsons model of politeness. However, their framework remains the foun-dation of politeness theory, and the list of output strategies they provide is theonly one to date which is sufciently detailed to allow for a ne-grained analysisof interaction data which takes adequate account of the local discourse context.In this instance, we have therefore chosen to follow their model in analysing the

interaction from the perspective of the speakers maintenance of the hearersface needs.

I L L U S TR ATI V E A N A LY S I S (3 ) O F I N TE R AC T IO NOur analysis focuses on the negative and positive politeness output strategiesused in the extract. A small selection of these is discussed and illustrated next.

Brown and Levinson list ten possible negative politeness output strategies,which are designed to address the hearers negative face needs. Claire makes



15/38

extensive use of output strategies from this category throughout the extract,especially in contexts where she is engaged in FTAs such as complaining, and dis-agreeing with or criticizing Toms statements or actions (see for example #1#5,#16, #44##49, #69#76, #86##89, #105##106, #123##133). Wecan see how Claire uses two of these strategies, minimize imposition and give defer-ence, in the following utterances near the beginning of the extract:

well i- the decision to make jared acting manager . . .well ive been overlooked quite a few times but I wanted to nd out specically

how what i could do to help myself be considered next time . . .(well) i just want to talk to you about it and i suppose i just want to get some

ideas on what i could do to actually be considered favourably next time . . .

Brown and Levinsons negative politeness output strategy minimize imposition isevident here in Claires repeated use of hedging phrases: well , i suppose and i just want to (which she repeats twice). Her use of the rst person pronoun in #5and #16: . . . what i could do . . . functions similarly by implying that Claire issimply seeking advice, not asking Tom to do anything about the situation him-self. This indirectness is also evident in #1 in her reference to the decision ratherthan describing it as Toms decision (as he in fact does himself in #33 so i took aunilateral decision) which would have immediately increased the degree of imposition on Toms negative face.

The negative politeness output strategy give deference is also present here. Inwording this as a request for advice Claire is acknowledging Tom as her superior,and positioning herself as a supplicant. She uses this strategy again in #106you- didnt think i had enough experience to act as manager i suppose, and in#133 so next time . . . you would consider me as the same as jared.

Brown and Levinson note the following possible motivations for use of thesetwo output strategies: (i) to give redress to the hearers want to be unimpingedupon; (ii) to avoid coercing the hearer; (iii) to minimize threat; and (iv) to makeexplicit power, distance and rank of imposition values. A politeness approach toanalysing the utterances above focuses on what Claire as the speaker is doing tomaintain her hearers face needs in this situation. This contrasts with the CDAanalysis by Marra and Holmes below where utterances such as #16 are seen as areection of Tom reinforcing the status difference and Claires role as suppli-cant, because he has made her repeat her reason for coming to see him, thusemphasizing the strategies used by the most powerful participant to maintain hisown face.

Brown and Levinson (1987) identify fteen output strategies for positivepoliteness. They note that these output strategies involve redress to the hearerswants (speaker wants hearers wants) (1987: 102), and therefore help to main-tain the hearers positive face. In contrast to Claire, Tom makes extensive use of positive politeness output strategies. One example is the way in which he givesreasons for what has taken place throughout the interaction. The followingutterances illustrate his use of this strategy:



16/38

yeah i dont think its a question of er favourability i mean it was a ques-tion more practicalities more than anything else um i was in urgent needof someone to ll in and jared had done that in the past already

so it wasnt a judgement call on were you better or he . . . better- it wassimply i saw precedents and that was the safest course of action in theshort time i had . . .

By saying that he took the actions he did, not because he did not want to see herin the job, but for a range of other reasons, he conveys that he wants Claireswants (cf. the CDA and discursive psychology analyses of these utterances byMarra and Holmes and Weatherall and Lane below, which place a rather differ-ent interpretation on Toms use of this strategy).

In the utterances listed below, Tom also expresses approval of Claire, thusreinforcing her positive face another positive politeness strategy:

. . . and ive got the report here yeah . . . looks very good i wouldve had no difculties in in um er acting you into the position . . . . . . you know um had I probably thought about it or um had this

conversation with you befo- i wouldve been quite happy . . . i mean next time it happens and if it does happen again then yeah sure

no difculties

Tom can also be seen to be using the positive politeness output strategy give giftsto the hearer (goods, sympathy, understanding, cooperation). His advice thatClaire should take the matter up with Joseph is also a particularly good exampleof how he uses this strategy:

i think thats a fair comment er i . . . personally would suggest that youknow you might like to raise that as a development issue with joseph . . .because hes your immediate controlling ofcer and i um you know i thinkhe should give you an opportunity um you know and and and or certainlytalk you through it

In this utterance Tom indicates explicitly that he is sympathetic to and under-stands Claires situation, and that he would like to nd a way to help her, i.e. thathe wants Claires wants. His repeated use of the addressee-oriented discoursemarker you know also helps to imply a sense of solidarity and shared under-standing.

To sum up, some FTAs threaten both positive and negative face. Brown andLevinson note that redressive action may be addressed to any potential aspect of the face threat (1987: 286). They also argue that negative politeness is morepolite than positive politeness. We have already seen how at the beginning of theextract, Claire is being as polite as she can be by using superstrategy four andgoing off-record in making a complaint (if we regard her as making a complaint).We can also see how she looks after Toms face needs throughout the extract byconsistently making much greater use of negative than positive politeness outputstrategies. Thus, according to Brown and Levinsons ordering, Claire is beingmaximally polite, both in her choice of superstrategy and in her choice of specic



17/38

output strategies, short of not making a complaint at all. This is consistent withthe status relationship between Tom and Claire. As Tom is Claires superior, wewould expect her to be particularly concerned to show deference and minimizethe imposition of an FTA if she cannot, or does not wish to, avoid enacting italtogether.

On the other hand, as illustrated earlier, Tom uses many more positive polite-ness output strategies throughout this interaction, both in comparison withClaire and relative to the number of negative politeness output strategies he uses.This pattern can also be interpreted to reect the differences between the two inpower and status. As a senior manager talking to a subordinate, Tom can affordto reduce the degree of social distance by means of a range of positive politenessoutput strategies, i.e. to be less polite. He also has a need to do so in the interests

of adequately defending or refuting Claires initial complaint, and in the interestsof maintaining good relations with her.Clearly, the hierarchical relationship between Tom and Claire is an important

factor in accounting for which politeness strategies each selects throughout thisinteraction, along with the nature of the FTAs being performed and the contextof the utterance. Although a politeness theory model does not explicitly accountfor or focus on the construction of status and power relationships, like the otherapproaches discussed here, it does assume intentionality on the part of thespeaker, and it also allows for the ranking of imposition on the hearers face of a given FTA. As this analysis has shown, it is therefore possible to address theissue of organizational power relations indirectly within this framework (seealso Morand, 1996), especially if sufcient weight is given the interactional and

social context in which an utterance occurs.

4. Critical discourse analysis (Meredith Marra and Janet Holmes)Nikander (1995: 6) describes discourse analysis as an umbrella for variousapproaches with different theoretical origins and therefore different analyticpunctuations and levels of analysis. Within this range, critical discourse analy-sis (CDA) is distinguished by its critical focus, its broad scope, and its overtlypolitical agenda (Kress, 1990: 845). As Titscher et al. point out, CDA sees itself as politically involved research with an emancipatory requirement: it seeks tohave an effect on social practice and social relationships (2000: 147).Consequently, CDA aims to reveal connections between language, power and ide-

ology, and critical discourse analysts aim to describe the way power and domi-nance are produced and reproduced in social practice through the discoursestructures of generally unremarkable interactions. They aim to raise peoplesawareness of the reciprocal inuences of language and social structure.

CDA has its theoretical roots in the work of Marx, Hall, Habermas andFoucault. At its core one nds investigations of the enactment, exploitation, andabuse of social power in everyday interactions. It is therefore particularly usefulin analysing interactions in settings involving a power asymmetry. Critical



18/38

discourse analysts typically take a top-down approach, examining the ways inwhich the superior controls or manages the discourse, using strategies based ontaken-for-granted assumptions about rights and obligations.

The precise analytical techniques used by different analysts to identify the rel-evant linguistic features of the text are very wide-ranging. Some focus on macro-level discourse strategies, examining rhetorical patterns, for example, whileothers adopt a conversation analytic or interactionally oriented approach; stillothers take a more grammatical approach, exploring relevant details of syntacticand semantic organization. Even at this level there is further diversity, with differ-ent critical discourse analysts using different grammatical models. So whilescholars adopting a CDA approach are committed to studying discursive forms of the use and abuse of power, the precise methods they use to achieve this are very

variable. In the analysis which follows, we illustrate ways in which a CDAapproach can illuminate the enactment of power in interaction. Like many CDAanalysts, we are relatively eclectic in the linguistic and discourse analysis tech-niques we use, but in this short illustration we have focused on just three featuresof interaction which we can illustrate from this particular text.

I L L U S TR ATI V E A N A LY S I S (4 ) O F I N TE R AC T IO NAdopting a CDA approach to the focus interaction led us to examine the strate-gies used by Tom to maintain control of the discourse in his interaction withClaire. A detailed example of how to apply CDA to spoken data is provided by vanDijk (1998). Here space restrictions mean we have limited discussion to just three of the many potential dimensions of analysis. These are: (i) the inuence of the setting

on one specic communicative act (interruption); (ii) consideration of one partic-ular argument structure ; (iii) consideration of one specic speech act ( summary ).Tom and Claire work in an organization characterized by hierarchical rela-

tionships. There are large differences in the degree of power and authoritywielded by different members of the organization. In this particular interaction,Toms role as a senior manager compared to Claires less inuential role as seniorpolicy analyst is evident in a number of features of the discourse, including thethree selected for analysis. (See the analysis using politeness theory by Vine andStubbe above for a complementary consideration of Claires discourse strategies.)

Setting The meeting takes place in Toms ofce, a space he owns. This has anumber of consequences for the structure of the discourse. We here discuss just

one of these, namely, the way that Tom deals with an interruption. First, Tomcan choose whether to accept or to ignore an interruption to the interaction. Bycomparison, if the focus interaction had taken place in Claires ofce, one wouldnot expect her to accept any interruption to her interaction with Tom. Instead of leaving the phone to ring, Tom interrupts Claires utterance with a (rhetorical?)request for permission to answer the phone, as he lifts the receiver:

can i just grab th- just grab that phone



19/38

Accepting an interruption from someone outside the interaction places onesaddressee in a one-down position. In this interaction, then, Tom has the power tohalt the conversation in which he is engaged with Claire in order to attend toanother addressee, while she has no right to act in this way, nor to object to histemporary demotion of her status as his addressee.

Claire does not object to the interruption, nor refuse Toms request. Rather,she explicitly accepts Toms apology (which is an indication that he recognizesthe negative function of the interruption in the interaction), with the response#9: thats okay, thus colluding with his marginalization of their interaction,and his clear relegation of it to a place of lesser importance than the interactionwith his phone addressee. (See also CA analysis by Lane and Hilder earlier.)

One consequence of the interruption is that Claire is forced to repeat her

reason for requesting the interview. Claires initial statement provided a varietyof linguistic indications that she was presenting herself as a subordinate seekingadvice: e.g. hesitations, repetitions, hedges (indicated in bold):

yeah um yeah i want to talk to you about um oh its a personal issue um+ well i- the decision to make um jared acting manager while joseph is away

Having to repeat a rather difcult and perhaps deliberately ambiguous speech act(i.e. presented as seeking advice but with a clear underlying potential for inter-pretation as complaint) puts Claire in a demeaning position. The repetition indi-cates a slight shift in emphasis towards seeking advice:

(well) i just want to talk to you about it and and i suppose [swallows][tut] i just want to get some ideas on what i could do to actually be consid-

ered favourably next timeBy requiring her to repeat and elaborate her reason for seeking an interview, Tomreinforces the status difference and Claires role as supplicant.

Argument structure and perspective Tom uses a wide variety of different types of argument in this interaction to justify the behaviour Claire is criticizing: onegroup of arguments relates to who has the right to dene the situation, i.e.whose perspective or version of what happened carries most weight:

1. Tom asserts his denition of the situation (#18#19, #31).2. He asserts his status and right to dene the situation (e.g. #126).3. He denies Claires denition (or at least his interpretation of her denition,

i.e. the denition he attributes to her) of the situation (#17, #29, #40, #62).(This point is discussed further in the CA analysis by Lane and Hilder earlier.)Toms approach involves asserting his interpretation and denying the valid-ity of Claires interpretation.

Another argument defends Toms behaviour as providing the simplest/most ef-cient solution (e.g. #31, #39, #40). (See discussion of practical considerationsin the discursive psychology analysis later by Weatherall and Lane.)



20/38

A third set of arguments constitutes an appeal to the power of the status quoin determining how to proceed. This takes the form of two closely related lines of argument: (i) an appeal to proper procedures or precedent; and (ii) an assertionof the inherent rights associated with the existing status hierarchy. We willfocus on these arguments.

In the rst line of argument, Tom repeatedly appeals to precedent as an argu-ment to justify his behaviour and support his advice to Claire:

um i was urgent need of someone to ll in and jared had done that in the pastalready

. . . it was simply i saw precedents [drawls]: and: that was the safest course of action . . .

. . . it was simply going on what was the safest ground in respect of what them- policy manager had done in the past

. . . in lieu of a decision ill take probably the last decision that was made. . . im more prone to take the least path of resistance or the path thats more

known to me which which which really was joseph had set a precedent before. . . (well as i say) i didnt er qualify my decision other than look at the

precedent

Appealing to precedent is a very conservative response to Claires concerns, onethat assumes and emphasizes the inherent rightness of the status quo. As theargument is elaborated by Tom, the word precedent, and its derivatives and syn-onyms, are often closely collocated with the words safe and safest (e.g. #40,#55). By using such arguments, Tom reinforces the existing power structure andorganizational hierarchy, rather than questioning and challenging them, analternative that a CDA approach highlights. CDA exposes the underlying taken-for-granted assumptions that enable power structures to recreate themselves andremain unchanged and unchallenged. Toms appeal to the safest procedures,the sensible, tried and true methods of dealing with a situation, namely to prece-dent, is a paradigmatic example of the way power relationships are performedand repeatedly reconstructed. Claires challenge is not welcomed as an opportu-nity to alter the status quo, or to question existing hierarchical relationships.Rather she is rmly re-placed in her subordinate position and told to follow theestablished rules if she wants to make progress in the organization.

In a second, closely related argument Tom asserts the importance of Claireusing the proper channels to make her request for consideration for preferment:

. . . the issue . . . is [drawls]: probably: one that um + you could address directly

with joseph. . . you might like to raise that as a development issue with joseph

He emphasizes his point by explicitly referring to Josephs status as Claires con-trolling ofcer:

. . . because hes your immediate controlling ofcer . . .

At a later point, he clearly stresses the rights of those with superior status:



21/38

. . . i always have the overriding nal say

Toms advice further reinforces his argument for abiding by the status quo. Hesuggests Claire takes a conciliatory approach to her superior, emphasizing hersubordinate status and her need to learn and obtain more training (#91#95).Moreover, at several points during the discussion, Tom refers to the way he him-self follows proper procedures in dealing with those of different status in theorganization:

. . . there would be very little chance of me crossing paths with the p m thepolicy manager . . .

um but ill never override my policy manager unless i thought it absolutelynecessary to do that and that would be quite rare

Thus Tom takes a management perspective on the issue Claire presents. He con-sistently appeals to proper procedures and asserts the importance of using thecorrect channels, namely those which the organization provides to deal with thesituation under discussion. His arguments presume the legitimacy of the existinghierarchical relationships, and take it for granted that Claire should act in anappropriately deferential manner in her dealings with her superiors.

Summarizing As a nal example of the way that Tom does power in this interac-tion, we point to the way he provides a summary of the argument at variouspoints, providing his denition or account of what has been discussed anddecided. It is also interesting to note, in passing, a closely related speech act,namely, agenda setting: Tom sets Claires agenda for future action.

Summarizing has been clearly identied as a strategy adopted by those in aposition of authority in workplace contexts to assist them in asserting and main-taining control of an interaction (see Sollitt-Morris, 1996; Holmes et al., 1999).Managers in meetings, for example, regularly summarize progress and set theagenda for future action (Holmes, 2000). There are a number of instances of Toms use of this strategy in the interaction under examination.

At a relatively early point, Tom summarizes the discussion, thus providing hisgloss or version of what has been agreed:

and it was as simple as thatso it wasnt a judgement call on were you better or he w- he better i- it was

//simply\ i saw precedents [drawls]: and: that was the safest course of action inthe short time i had

/(right)\\At the end of the discussion, Tom again summarizes the interaction with . . . thats really what it boils down to . . ., and with a fuller summaryof his defence in line #146. He concludes: so (now-) i mean + nexttime it happens and if it does happen again then yeah sure no difculties. Thissummary expresses a superior, patronizing position in relation to Claire. Tomconstructs his role as a reassuring adviser to an acolyte in need of advice.



22/38

Correspondingly, at several points, Tom outlines future steps for Claire to take(e.g. #79, #81#82, #91, #95) in order to avoid a repetition of the situation.To sum up, our analysis indicates three different ways in which power and

dominance are expressed, enacted and reproduced in discourse, both in its struc-ture and its contents, three different ways in which discourse reproducesinequality (van Dijk, 1999: 460).

We have illustrated how:

1. a superiors management of one discourse feature (interruption) in oneparticular setting can reinforce the existing power relationship betweeninteractants;

2. particular argument structures may support the existing power relation-ships, and imply the given and incontestable nature of those relationships;

3. specic speech acts such as summarising (and the related act of agendasetting) can be used to emphasize and reinforce the dominant persons deni-tion of the situation and the relationships.

5. Discursive psychology (Ann Weatherall and Chris Lane)Discursive psychology has emerged since the late 1980s as part of what has beenidentied as a turn to language across the humanities and social sciences (Gill,1993). An important aspect of discursive psychology is that it has developedwithin its parent discipline in conjunction with critiques of conventionalresearch practices, such as experimentation and quantication, and a question-ing of the dominant epistemological assumptions of realism and positivism. Thusdiscursive psychology is not so much a method as a theoretically informed ana-lytic approach for understanding social psychological phenomena such as iden-tity, inter-personal and inter-group relationships, persuasion, discrimination andprejudice (see Potter and Wetherell, 1987; Edwards and Potter, 1992, foroverviews).

It is possible to distinguish between different styles of discourse analysiswithin discursive psychology. At a general level, they are all more or less inu-enced by linguistic philosophy and pragmatics, ethnomethodology, conversationanalysis and post-structuralism. The style varies depending on the theoreticalemphasis; however, key concerns are the practices and resources available to jus-tify, rationalize and guide social conduct. One of the analytic aims of discursivepsychology is to examine the (linguistic and discursive) resources that are used torationalize and justify social practices (for example gender inequality). A broaderaim of some styles of discourse analysis may be to consider how patterns of lan-guage use, sometimes referred to as interpretative repertoires, practical ideolo-gies or discourses, function to recreate and sustain wider patterns of socialinequality.

It is simply not possible in a short amount of space to give a more comprehen-sive description of the development and nuances of what has become a broad



23/38

and complex approach in social psychology. Hopefully, the general avour andvalue of a discursive psychological approach to analysing talk-in-interaction arecaptured through the sample analysis.

I L L U S TR ATI V E A N A LY S I S (5 ) O F I N TE R AC T IO NThe selected interaction provides an excellent extract for analysis using a discur-sive psychology approach. It is a naturally occurring, spontaneous interactionconcerning employment opportunities. Also, there is a dilemma of stake orinterest in the interaction. Tom must account for the decision to make Jaredmanager at the same time as presenting himself and the decision as reasonableand rational. How is that dilemma handled? One of the questions a discursivepsychologist would ask is: how does Tom present himself as blameless in the deci-

sion to appoint Jared as manager? Also, what explanations for the decision doesTom use to justify it? However, the rst stage of the analysis involves establishingthat there is some kind of dilemma of stake being managed in the interaction.

Establishing the dilemma of stake In #1 and #5 Claire raises the issue that shewishes to discuss with Tom. That issue is the decision to make Jared rather thanher acting manager. Claire frames the issue as personal and her motivation forraising the issue as to nd out . . . what I could do to help myself be considerednext time. In #16 she reiterates that she wants to get some ideas on what Icould do to actually be considered favourably next time. There is some ambiguityassociated with Claires utterances in #1, #5 and #16. They could be interpretedas a request for advice or (and perhaps as well as) a complaint about discrimina-

tion.Toms rst response in #17 I dont think its a . . . question of favourabilitydemonstrates he is orientating to Claires utterance as a complaint about discrim-ination rather than as a request for advice. The spontaneous denial in #62absolutely nothing sinister or any other agenda provides further evidence that arelevant aspect of the interaction for Tom is the complaint about bias. Later inthe interaction (around #74) the orientation shifts to addressing the request foradvice. However, here the focus is on the aspects of the interaction that attend tomanaging what Tom orientates to as a complaint.

Previous discursive research has documented a widespread tendency forpeople to deny personal bias. The reluctance to articulate explicitly prejudicialattitudes has led to the suggestion that there are cultural norms against expres-

sions of, for example, racism or sexism. Thus, as already mentioned, an aim of the analysis is to examine how inequality gets articulated, rationalized and justi-ed. Given that the issue of bias is a relevant aspect of the interaction, then thenext step of the analysis is to identify the resources and devices used to accountfor the decision.

Before proceeding, it is useful to highlight some features of the analyticprocess so far that characterize it as discursive psychology. First is the notionof truth. The analysis is not concerned with establishing whether or not the



24/38

decision to make Jared manager was really an instance of discrimination. Rather,as in CA, the important point is that it is evident from the conversation itself thatTom is orientating to the issue of bias as relevant. Further, the analysis is notconcerned with measuring Toms attitudes or beliefs that may have led him to abiased decision. Rather, the focus is to document how the topic of inequality getsmanaged in talk in ways that make it seem rational and fair.

Practices and resources Given that Tom is concerned with presenting himself asfair minded, and the decision about appointing an acting manager as reasonable,what linguistic resources are used to that end? Previous research on the discus-sion of employment opportunities and equity issues has made impressiveprogress in documenting the discursive patterning of bias (see Wetherell et al.,

1987; Gill, 1993; Gough, 1998). One resource that has been identied and veri-ed as functioning to naturalize and justify inequality is practical considerationtalk. In the workplace example, reference to practical considerations was used toaccount for the decision at the following points of the interaction:

i mean it was a question more practicalities more than anything else. . . it was simply logistics and what was practically easy that would create the

least amount of hassles . . .

Tom constructs tradition or precedents and time pressures, not as potentially per-petuating inequality, but as practical considerations that justied the decision.References to precedents and time pressures as part of the practical considera-tions were:

um i was urgent need of someone to ll in and jared had done that in the pastalready

. . . i saw precedents and that was the safest course of action in the short timei had

. . . it was simply going on what was the safest ground in respect of what thepolicy manager had done in the past

. . . in lieu of a decision ill take probably the last decision that was madel- l- let me just say i im more prone to take the least path of resistance or the

path thats more known to me which which which really was joseph had set aprecedent before

. . . i didnt er qualify my decision other than look at the precedent

The emphasis that Tom gives to practical considerations serves to construct him-self as blameless and helps justify an arguably unfair decision as reasonable andrational.

A second resource that has been identied as accounting for inequality is thenotion of difference (see Gill, 1993; Gough, 1998). In contrast to practical con-siderations talk that denies the inuence of personal characteristics in the deci-sion, difference talk justies the outcome on the basis of comparisons betweenindividuals. Differences in knowledge, experience or skills may be used to justifyinjustice. So despite Tom claiming in #29 and again in #37 that the decision was



25/38

not based on any comparison between Claires and Jareds capabilities, differencetalk was used at the following points of the interaction to justify the decision:

. . . jared had done that in the past already. . . it had a little bit also to do with the fact that er a number of very current

issues that i had been involved with jared on er like [name of organization] +and like the [topic] er issue

um i probably had more immediate contact with him you knowum and it was just because + jared was more um currentand also the fact that he had been put in there beforejareds got um [drawls]: er: you know obviously hes been working in regional

councils and things like that so hes. . . on the eld side hes probably got a fair bit of experience

The notions of practical considerations and differences were two of the resourcesin Toms talk that constructed the decision as fair rather than biased. A furtherstrategy that Tom uses is to deny or downplay his personal role in a decision thatcould be construed as biased:

. . . there would be little chance of me crossing paths with the p m the policymanager . . . for his reasons

you know joseph hadnt declared who he wanting acting up. . . . . .. . . he [joseph] didnt articulate that [a different decision] and um i always

have the overriding nal say. . . but ill never override my policy manager unless i thought it absolutely

necessary

The practical considerations and difference talk, combined with Toms down-

playing of his role in the decision to promote Jared, form a compelling accountthat justies and rationalizes (what Tom orientates to as) a biased decision.Furthermore Tom constructs himself as having little agency in Jareds promo-tion. Rather, precedents, time and Jareds unique skills are the reasons for thedecision. Indeed, Tom (in #79#81) implies that it may be Claires fault for notaddressing the issue directly with her controlling ofcer, even though he con-cedes that what she is raising is quite valid (#135) and that he would considerher favourably next time (#133#134 and #151). Thus Tom constructs bothhimself and the decision as reasonable and fair.

In summary, discursive psychology, unlike more linguistically informedapproaches to discourse analysis, does not aim to give a better account of lan-guage structure. Rather discursive psychology documents the strategic variabil-

ity, construction and functions of language. The focus is on the broader patternsof meaning making that are resources for social actions. The present analysis hasexplicated some important ways in which certain ideas can be used to reinforceand justify bias. In this way, discursive psychology can contribute to other formsof discourse analysis that aim to document the discursive (and material) repro-duction of social inequality such as racism and sexism.



26/38

DiscussionThe ve approaches to analysing language in use outlined above all differ in theirtheoretical orientation and analytic approach to talk and text, and thereforepresent some rather different perspectives on the sample interaction. Despite thisdiversity, however, a number of common themes do emerge from each analysis,as indeed we would expect. For example, at the most basic level, each analysisdeals in some way with how Claire presents her complaint/request and/or howTom responds, and treats this as an example of problematic or conictual dis-course. However, the precise nature of the problem is unpacked from differentangles, ranging from the more micro-analytic approaches such as conversationanalysis, interactional sociolinguistics and politeness theory, through to discur-sive psychology and critical discourse analysis which focus more on the linguisticrealizations of broader meaning systems or discourses.

With the possible exception of CA, all ve approaches consider the interactionbetween language and social structures, but draw on the situational context andbroader socio-cultural factors to inform the analysis more or less explicitly, and todifferent degrees. For example, the status/power asymmetry between Clare andTom emerges from all the analyses, as do aspects of the talk that are seen as(re)producing dominant socio-cultural practices.

A further element common to all the analyses is that each identies a range of linguistic and discoursal strategies or forms that participants select from, againat a number of different levels, in order to achieve particular goals. Each analysisalso recognizes that meaning resides in the interaction of linguistic form andsocial context, and that utterances are in themselves intrinsically indeterminate.It is therefore possible for the participants (or the analyst) to assign a number of different interpretations or readings to a given utterance or sequence. The com-plex interplay between different meanings and interpretations (both actual orpossible), and the strategic use made of this by the participants is explored from anumber of different angles in the various analyses presented here: e.g. maintain-ing the addressees face in politeness theory, doing power in CDA, the use of framing and contextualization cues in IS, and the use of conversationalresources by participants to invoke or orient to particular interpretations in CAand discursive psychology .

However, the sample analyses presented in this article show that, althoughthere are elements common to all the analyses, each perspective clearly has its

own unique features, and highlights different aspects or dimensions of the inter-action. These distinctive elements are summarized next.

C O N V ER S AT I O N A N A LY S IS5

Conversation analysis provides a strict empirical framework for analysing indetail the way participants jointly construct the interaction and at the same timeconstitute the context, including participants identities, utterance by utterance.It shares a focus on contextualization processes with interactional sociolinguis-



27/38

tics, and an interest in accounting and justifying practices with CDA and discur-sive psychology. However, CA more than any other approach insists on treatingas relevant to the analysis only that which participants themselves display as rel-evant in the interaction itself. This is perhaps best exemplied by the ways inwhich an analyst working strictly within a CA framework would take issue withseveral aspects of the other analyses.

One example is the assumption in the other analyses that the decision toappoint Jared was Toms. Tom certainly takes responsibility for it, but whose deci-sion it was (and whose it should have been) is not entirely clear-cut from theinteraction itself, as shown by Claires contributions at #44#45 i suppose itsbecause um joseph hadnt really talked to me beforehand why hed chosen jaredbecause hed only been there a short while and #123#124 well actually

joseph had decided (we both) would be . . . acting (as managers) he came in onFriday and said that (both seniors) would be um acting. Particularly at thebeginning of the extract, a CA analyst would hold that we cannot assume thatClaire is holding Tom responsible for the decision. This position contrasts withthat taken in the IS analysis, which admits contextual information gathered sep-arately from the recording as relevant to the analysis, and is therefore able tocome to a more denite conclusion.

For similar reasons, a conversation analyst would also take issue with thecognitivist approach taken by politeness theory, which assumes that we canattribute certain motivations and intentions to speakers simply because they usea particular strategy. An important question here is: what is the action thatClaire is initially engaged in, and what can be discounted as being merely a real-

ization of a politeness strategy? And how do you decide? For example, it is tempt-ing to regard Claires initial request for advice/mentoring as a cover for acomplaint, but looking later in the interaction Claire appears to resist Tomsresponding to it as a complaint, while she repeats and pursues her request foradvice. Perhaps that is a continuation of the same politeness strategy, but per-haps she really means it.

Another point of difference between CA and some of the other approacheslies in CAs focus on interaction as a joint activity. One can ask in relation to thediscursive psychology analysis, for instance, why it is that Tom gives such anextensive justication of the decision and his position. From a CA point of view,this comes about through a joint process which includes Claire refraining fromoffering positive acceptance of Toms explanations and advice, and her persist-

ence in raising the issue of what happens next time. If Claire had offered a posi-tive acceptance of Toms rst explanation, that might have been as far as Toms justication went.

CDA provides the strongest contrast to CA. There are at least three major criti-cisms of the CDA analysis which conversation analysts could make: (i) theCDA analysis is based on particular analyst readings of the interaction, withoutreference to the participants own readings, and with no obvious justicationfor excluding other plausible readings; (ii) the analysis is largely based on the



28/38

analysis of items isolated from their sequential context; and (iii) the analysisappears to be circular, in that it assumes the pervasive relevance of dominanceand power, and then claims to show how patterns of dominance and power per-vade normal everyday talk.

I N T E RA C T IO N AL S O C I OL I N G UI S T I CS

Interactional sociolinguistics focuses explicitly on the social and linguistic mean-ings created during interaction, and like CA, models interaction primarily as jointactivity, although IS assumes somewhat more intentionality on the part of thespeaker than would be the case in CA. Participants are assumed to make situatedinferences about one anothers communicative intentions and goals based on awide array of verbal and non-verbal cues that form part of cultural repertoires

for signalling meaning. From the analysts point of view, this involves combiningan analysis of the unfolding conversation as co-constructed joint activity with aknowledge of the wider sociocultural context to understand how the discourseindexes pre-existing socio-cultural meanings. Contextualization cues are seen toindex an intended meaning or activity, on the basis of shared knowledge of soci-olinguistic or cultural norms. Like CDA and discursive psychology, interactionalsociolinguistics is also often applied to a critical analysis of texts from the per-spective of power differentials, especially in cross-cultural communication.

P O L I TE N E S S T H E O RY

Politeness theory links a focus on speaker/hearer intention and interpretation toa sociological and anthropological concern with how interaction relates to social

structures. It examines how speakers protect and/or threaten the face of theirinterlocutor through the use of specic superstrategies and output strategies. Incontrast to CA and IS, which model interaction as a joint construction, politenesstheory is based on a transmission model of interaction, taken from the perspec-tive of the speaker. In this approach, utterances are typically analysed one at atime, with the emphasis on identifying the specic output strategies. There is alsoa tendency to attribute particular intentions to the speaker, and in so doing,assigning only one possible interpretation of the utterance. Although the theorydoes allow in principle for the possibility of alternative interpretations, the waythe framework is set out makes it more difcult to address this fact in any depth.Similarly, issues of power are addressed only indirectly in the Brown andLevinson framework, as an explanation for the selection of particular strategies.

C R IT I CA L D I SC O UR S E A NA LY SI S

Critical discourse analysts, like others working within a social constructionistframework, see discourse as a form of social practice. Critical discourse theoryfocuses explicitly on exploring how power and ideology are manifested in dis-course, and on the linguistic aspects of social and cultural processes and struc-tures. CA, IS and politeness theory all, at least in part, take a bottom-upapproach which is grounded in a turn-by-turn analysis of the interaction from



29/38

the perspective of each participant in turn. CDA, on the other hand, like discur-sive psychology, takes a much more top-down and extensive perspective in twoways: (i) it provides a more global view of the types of linguistic forms/strate-gies used; and (ii) it takes the perspective of the most powerful participant. Its aimis to reveal connections between language, power and ideology how are powerand dominance (re)produced by means of discourse structures?

Leading proponents of CDA and CA have debated the relative strengths andweaknesses of their respective approaches in recent issues of Discourse and Society (Volumes 8 (2), 9 (3), and 10 (4)). The editor, van Dijk, summarizesSchegloffs criticisms of CDA as follows: CDA is often short on detailed, system-atic analysis of text or talk, as carried out in CA (1999: 459).

The CDA analysis above demonstrates, albeit briey, that this is not a neces-

sary consequence of adopting a CDA approach. CDA entails three stages of analysis: description of text, interpretation of the relationship between text andinteraction, and explanation of the relationship between interaction and socialcontext (Fairclough, 1989: 109). The description of the text can be just asdetailed as the analyst judges appropriate and necessary to expose the underly-ing ideological assumptions and power relationships of the participants.

Critical discourse analysts explicitly adopt a particular political viewpoint,one which aims to expose and uncover taken-for-granted power relationships.By illustrating how patterns of dominance and power pervade normal every-day talk, the analyst draws attention to issues of inequality in society, with thepotential for bringing about change. This is by no means a simple task, as vanDijk indicates:

Critical discourse analysis is far from easy. In my opinion it is by far the toughest chal-lenge in the discipline. . . . [I]t requires true multidisciplinarity, and an account of intricate relationships between text, talk, social cognition, power, society and culture.(1998: 370)

D I S C UR S I V E P S Y CH O L O G Y

Discursive psychology focuses on the discursive practices and resources availableto justify, rationalize and guide social conduct. In this case it was used to high-light how patterns of language use (discourses, interpretative repertoires, practi-cal ideologies) maintain and reproduce social inequality. This approach operatesat more of a macro-level than some of the other frameworks. In its emphasis onrecurring themes and argument structures, it focuses more on discourse

processes and linguistic forms as a vehicle for research into social psychologicalissues such as identity and interpersonal or intergroup relationships languageand communication are seen as the site of the social rather than these beingthe target of the analysis in their own right. However, it shares with CA andIS an interest in how themes are contextualized and constructed through thediscourse. On the other hand, although discursive psychology has much incommon with CDA, most notably its relatively macro-level and more extensiveapproach to analysis and its critical focus, it nevertheless has a somewhat



30/38

different if overlapping area of interest. For example, discursive psychologistsoften look at other sociopsychological phenomena, e.g. memory, as well as issueslike the social construction of power and discrimination.

Conclusion

Although there are substantial areas of overlap, as we have seen, there are also anumber of signicant differences in emphasis or perspective and some tensionsbetween the ve different approaches to discourse analysis presented here. First,each framework takes a slightly different approach to the place of extra-textualcontext in the analysis, ranging from the strong version of CA which claims notto make use of any information outside the local interactional context, through

to weaker versions of CA, IS and politeness theory, which do admit contextualand socio-cultural information to a greater or lesser extent, and then to CDA anddiscursive psychology which also include a focus on the broader socio-politicalcontext and existing social discourses, particularly those relating to power.

Second, there are differences in the level of detail with which linguistic, para-linguistic and discourse features are analysed. CA works primarily within amicro-analytic framework. IS and politeness theory operate at this level too, butare also concerned with identifying more general patterns (indexicality andsocio-cultural norms in the case of IS, superstrategies in the case of politenesstheory). CDA and discursive psychology both attend to the big picture in order toidentify the constructs which provide the underlying logic for the specic discoursestrategies that are used in an interaction, but CDA in particular can also, and often

does, accommodate a much more ne-grained analysis of relevant excerpts.A third difference concerns the degree to which an interaction is seen and/oranalysed as a joint construction, as opposed to the more traditional view of com-munication as a simple transmission of information or intent. Politeness theoryts most obviously into the latter category, while CA and IS take a strong andweak social constructionist approach respectively. The other approaches fallsomewhere in the middle. The contrast along this dimension is also reected indifferent sets of assumptions about intentionality and inferencing.

Each approach therefore provides a slightly different lens with which toexamine the same interaction, highlighting different aspects or dimensions of itskey features. These are not necessarily in conict with one another (though insome cases the analyses and/or the theoretical assumptions underlying them are

difcult to reconcile); rather, they are complementary in many ways, with eachapproach capable of generating its own useful insights into what is going on inthe interaction, with the proviso that the framework adopted needs to be a goodmatch for the research questions being asked. Hence, while the exercise we haveengaged in clearly presents many challenges, we hope this article has demon-strated the value of analysing one text from a range of perspectives, and theinsights to be gained by applying a range of different theoretical and method-ological approaches to the same piece of discourse.



31/38

A P P E N D I X 1 : T R A N S CR I P T IO N C O N V E NT I O N S

[laughs]: . . ..: Paralinguistic features, descriptive comments, time codes[name of section][10.00]+ Pause of up to one second. . . .1//. . .\.1. . . . .2//. . . .\2 Simultaneous speech. . .1/. . .

Multiple Discourse Analysis

Documents

Transcript of Multiple Discourse Analysis